Hacker News Reader

Comment by directevolve | Hacker News Reader

It would take a lot more than an old, low-confidence set of trials (which is how the meta-analysis describes its inputs) to convince me of your premise.

View on HN

The default position is that a treatment will do nothing to help for a condition. You need to see evidence for DBS helping with Parkinsons to believe it will, lack of strong evidence in favor of effectiveness just reverts to the default position.

So if that meta study finds that generally all evidence related to DBS effects on Parkinsons is low-confidence and old, then why believe that it is even the right direction?

I’m sure there is plenty of evidence for DBS being plausibly safe and effective when they get the technology right. It will be in the form of:

- Studies on the purported mechanism(s) of action

- Preclinical results in vivo

- Anecdotal results like the one in the BBC article

- Mixed effects from older versions of the tech

That’s plenty to motivate further research and trials into this treatment.

Again, we have a meta-study that has looked at all of the available evidence up to ~6 months ago, and has found no strong evidence that DBS is a promising treatment option. Better devices to achieve the un-promising treatment won't change that fact, not significantly.

Since Parkinsons is for now an incurable and barely treatable disease, it makes sense to try even things that are not very promising. But that doesn't mean we should hold out much hope for them, when all evidence suggests that they are not going to pan out.

Your points are all things that the authors of the meta-study should have looked for, or irrelevancies (anecdotes of people feeling better at some points after they underwent major neuro-surgery are entirely irrelevant to the complexity of a very slow acting degenerative disease).

Reply on news.ycombinator.com